Flash Distillation

One valve, one drum, one stage

Imagine a stream of hot liquid held at, say, 30 bar and 250 °C. At that pressure it stays liquid even though it is well above its atmospheric boiling point — pressure is the only thing keeping it from boiling. Now send it through a throttle valve into a vessel held at 3 bar. The instant the pressure drops, the saturation temperature for that mixture collapses, and the feed finds itself superheated relative to its new surroundings. It cannot stay all-liquid, so it does the only thing it can: it boils. A fraction of the feed flashes to vapor in a fraction of a second, and the two phases separate by gravity inside the drum — vapor rising to the top, liquid pooling at the bottom. That is the entire device. No trays, no packing, no condenser, no reflux. One throttling step, one phase split, one ideal equilibrium stage.

The point of doing this is separation. The vapor that flashes off is not a representative sample of the feed — it is preferentially loaded with the components that want to be vapor, the volatile ones. The liquid that stays behind is correspondingly concentrated in the heavy, less-volatile components. So a single flash takes one mixture and hands you two streams of different composition. It is the simplest non-trivial separator that exists, the limiting case of distillation when you allow yourself exactly one equilibrium contact.

Vapor–liquid equilibrium does the sorting

The reason the vapor and liquid have different compositions is vapor-liquid equilibrium (VLE). At equilibrium, each component i distributes between the phases according to its equilibrium ratio (K-value):

K_i = y_i / x_i

where y_i is the mole fraction in the vapor and x_i the mole fraction in the liquid. A component with K > 1 prefers the vapor; one with K < 1 prefers the liquid. For ideal mixtures, K_i follows from Raoult's law: K_i = P_i^sat(T) / P, the ratio of the pure-component saturation pressure (a strong function of temperature via the Clausius-Clapeyron relation) to the total drum pressure. That single line tells you everything that matters for design: raise the temperature or drop the pressure and K rises, sending more material to the vapor.

What actually controls how well the flash separates two components is their relative volatility:

α₁₂ = K₁ / K₂ = (P₁^sat / P₂^sat)

If α = 1 the two components have identical K-values, split identically, and the flash does nothing. The larger α is, the more lopsided the vapor enrichment. A flash separating propane (α ≈ 5–10 vs. heavier hydrocarbons) does real work in one stage; a flash on ethanol-water (α ≈ 2–3, falling to exactly 1 at the 95.6 wt% azeotrope) barely moves the needle. Relative volatility is the hard ceiling on what a single stage can achieve, no matter how you tune temperature and pressure.

The flash calculation

Designing a flash means solving three coupled relations for the feed of overall composition z_i: the component mass balance, the equilibrium relation, and the energy balance. Define the vapor fraction ψ = V/F. Combining the mass balance F·z_i = V·y_i + L·x_i with y_i = K_ix_i gives the per-component split:

x_i = z_i / [1 + ψ(K_i − 1)]    y_i = K_i z_i / [1 + ψ(K_i − 1)]

Because all the liquid mole fractions must sum to one and all the vapor mole fractions must too, you get the celebrated Rachford-Rice equation, a single scalar equation in the one unknown ψ:

Σ_i   z_i(K_i − 1) / [1 + ψ(K_i − 1)] = 0

Solve it (it is monotonic in ψ, so Newton's method converges in a handful of iterations) and you have the full split. A useful check: a physically meaningful two-phase flash exists only between the bubble point (Σ K_iz_i > 1) and the dew point (Σ z_i/K_i > 1). Outside that window the drum runs all-liquid or all-vapor and there is nothing to separate.

The energy balance is what actually fixes ψ. The flash is essentially adiabatic — the drum exchanges almost no heat with its surroundings on the timescale of the flash. The latent heat needed to vaporize ψ moles must come from somewhere, and the only reservoir is the sensible heat of the feed. So the stream cools itself as it boils, exactly the way an expanding gas cools in the Joule-Thomson effect — except here the cooling is dominated by latent heat rather than gas non-ideality. The drum temperature settles wherever the enthalpy balance H_feed = ψ·H_vapor + (1−ψ)·H_liquid closes:

ψ · Δh_vap ≈ c_p · (T_feed − T_drum)

For water, Δh_vap ≈ 2257 kJ/kg and c_p ≈ 4.2 kJ/(kg·K), so flashing seawater from 90 °C down to 40 °C — a 50 K drop — releases enough sensible heat to vaporize only about ψ ≈ 4.2 × 50 / 2257 ≈ 9% of the stream per stage. That tiny per-stage yield is precisely why desalination plants chain many flash chambers together.

Why the drum runs cold

The self-cooling is the most counter-intuitive part of the device for newcomers. You feed it hot liquid; the liquid that comes out the bottom is colder than what you put in, even though no heat was removed. The energy did not vanish — it was repackaged from sensible heat (temperature) into latent heat (the kinetic and potential energy of the molecules that escaped to the vapor). The remaining liquid sits exactly on the bubble curve at the drum pressure: it has cooled until it is just barely boiling and no further. If you draw the process on a phase diagram, the feed point lies in the superheated-liquid region for the low pressure, and the flash carries it horizontally onto the two-phase dome, splitting into a saturated-vapor point and a saturated-liquid point connected by a tie line whose lever-arm ratio is ψ.

Flash vs. fractional distillation

The defining limitation of a flash is that it gives you exactly one equilibrium stage. Real fractional distillation stacks dozens of stages in a column with internal reflux, multiplying the single-stage enrichment over and over until you reach high purity. The comparison below frames where each belongs:

Property	Flash distillation	Fractional (column) distillation
Equilibrium stages	1	10–100+ (trays or HETP units)
Reflux	None	Essential (sets purity/energy trade-off)
Typical purity (single pass)	Modest enrichment, set by α and ψ	Up to 99.9%+
Capital cost	Low — one drum, one valve	High — tall column, condenser, reboiler
Energy use	Very low (uses feed's own enthalpy)	High (reboiler duty plus condenser duty)
Best for	Easy splits, pre-concentration, knocking light ends off a feed	Tight splits, fuel-grade purity, close boilers
Footprint	Small, robust, no moving internals	Large, fouling-prone trays

In practice the two are partners, not rivals. A refinery crude unit preheats oil to ~360 °C and runs it through a flash zone at the bottom of the atmospheric column: the flash does the bulk, easy separation of light ends from residue for free, and the trays above it do the fine fractionation into naphtha, kerosene, and diesel. The flash is the cheap workhorse that takes the brunt of the load so the expensive column does not have to.

Equilibrium flash vs. differential (Rayleigh) distillation

It is worth distinguishing the flash from its close cousin, differential or batch (Rayleigh) distillation, because they sound similar but separate in opposite ways. In a flash, the entire vapor stays in intimate contact with all the liquid until they disengage together — one equilibrium contact between the whole vapor and whole liquid. In a Rayleigh still, each tiny increment of vapor is removed the instant it forms, so it equilibrates only with the momentary liquid and then leaves; the liquid composition drifts continuously. The flash gives a cleaner single-stage cut; the differential still integrates over a changing liquid and is described by the Rayleigh equation ln(L/L₀) = ∫ dx/(y−x). Both are single-stage in spirit, but the flash holds everything together while the Rayleigh process strips vapor away as it appears.

Feature	Equilibrium (flash)	Differential (Rayleigh)
Vapor contact	Whole vapor with whole liquid at once	Each vapor increment removed immediately
Operation	Continuous, steady-state	Usually batch, time-varying
Governing math	Rachford-Rice (algebraic)	Rayleigh equation (integral)
Composition over run	Constant once steady	Drifts as light ends deplete

Where flash distillation earns its keep

• Multi-stage flash (MSF) desalination. Hot seawater (~110 °C) cascades through 15–25 chambers, each at successively lower pressure. A few percent flashes in each stage; the cumulative distillate supplies a large share of the freshwater in the Persian Gulf, with the world's biggest plants making well over a billion liters per day. Recovered latent heat preheats incoming brine, making the process surprisingly energy-frugal per liter.
• Crude-oil preheat / flash zone. The flash zone at the base of an atmospheric distillation tower vaporizes the light and middle distillates from heated crude before the trays fractionate them — the first separation every barrel of oil sees.
• Geothermal power. Single- and double-flash plants throttle hot pressurized brine to flash off steam that spins turbines; the leftover brine is reinjected. This is how a large fraction of geothermal electricity is generated.
• Refrigeration and economizers. A flash tank in a two-stage compression cycle separates flash gas from liquid refrigerant, feeding cool vapor to the second compressor stage and improving the coefficient of performance.
• Gas processing and solvent regeneration. Rich amine or physical solvent loaded with absorbed CO₂/H₂S/light hydrocarbons is flashed to a lower pressure to recover dissolved gases before the costly thermal regenerator.
• Steam flashing / blowdown heat recovery. Boiler blowdown at high pressure is flashed to make low-pressure steam, recovering energy that would otherwise be dumped.

The two knobs that matter

A flash designer has exactly two degrees of freedom once the feed is fixed: the drum pressure and the feed preheat temperature. Together they set ψ and therefore the split. Push the feed hotter or drop the drum pressure and you flash more vapor (higher ψ), pulling more total material overhead but also dragging some of the heavy key with it, eroding purity. Run cooler or at higher pressure and ψ shrinks toward a tiny, very-pure vapor cut. The sweet spot balances recovery (how much of the light key you capture overhead) against purity (how little heavy key contaminates it). Because there is no reflux to clean up the vapor, you cannot have both at once in a single flash — that fundamental ceiling is the reason flashes are paired with columns whenever tight specs are required. There is also a physical sizing constraint: the drum must be large enough that the upward vapor velocity stays below the entrainment limit (the Souders-Brown criterion), or liquid droplets get carried overhead and quietly ruin the separation.